Patch: mips opcodes: Fix lwxc1.

Fu, Chao-Ying fu@mips.com
Thu Jun 19 22:23:00 GMT 2008


Fu, Chao-Ying wrote: 
> 
>   Somehow I got a failed test for "odd-float.s" due to "..." 
> in "odd-float.d".
> I just wonder if this test can pass for you.  Or, do we need 
> to remove "..."
> in "odd-float.d"?  Thanks!
> 
> Index: odd-float.d
> ===================================================================
> RCS file: /cvs/src/src/gas/testsuite/gas/mips/odd-float.d,v
> retrieving revision 1.1
> diff -u -p -r1.1 odd-float.d
> --- odd-float.d 4 Oct 2007 21:53:06 -0000       1.1
> +++ odd-float.d 19 Jun 2008 20:51:00 -0000
> @@ -7,4 +7,3 @@
>  Disassembly of section .text:
>  0+00 <[^>]*> lwxc1     \$f1,a0\(a1\)
>  0+04 <[^>]*> swxc1     \$f3,a0\(a1\)
> -       ...
> 

  I think I know the problem.  Because I use elfoabi targets,
the section alignment isn't at 16-byte boundaries and
nops aren't generated.  So, instead of changing "odd-float.d",
we may change "odd-float.s" as follows.

Index: odd-float.s
===================================================================
RCS file: /cvs/src/src/gas/testsuite/gas/mips/odd-float.s,v
retrieving revision 1.1
diff -u -p -r1.1 odd-float.s
--- odd-float.s 4 Oct 2007 21:53:06 -0000       1.1
+++ odd-float.s 19 Jun 2008 21:47:08 -0000
@@ -4,3 +4,6 @@
 text_label:
        lwxc1   $f1,$4($5)
        swxc1   $f3,$4($5)
+
+# Force at least 8 (non-delay-slot) zero bytes, to make 'objdump' print ...
+       .space  8


  Thanks!

Regards,
Chao-ying



More information about the Binutils mailing list